How we investigate and score crypto casinos

Two outputs, not one opaque rating

The casino score runs from 0 to 100 and measures documented operating quality. The confidence grade measures how complete, current and corroborated the evidence is. A platform can therefore receive a respectable score with limited confidence, or a poor score supported by strong documentation.

Verdict bands remain consistent: 80–100 indicates strong legitimacy signals; 65–79 legitimate with limitations; 45–64 elevated caution; 25–44 high suspicion; and 0–24 critical risk. Closed products and migrated brands receive explicit status wording instead of a misleading live-casino verdict.

The ten weighted tests

The operating score is assembled from ten tests. Each criterion is capped at its published weight, so a large game catalogue cannot compensate for an unidentified operator or an unresolved withdrawal structure.

• Operator identity — 15 points
• Licence and regulatory scope — 15 points
• Withdrawals and KYC — 15 points
• Payments and crypto routing — 12 points
• Complaints and redress — 10 points
• Domain and account security — 8 points
• Game integrity and product reality — 8 points
• Terms and enforcement — 7 points
• Player-protection controls — 5 points
• Support continuity — 5 points

How criterion scores are assigned

A criterion starts with the strongest verifiable evidence: regulator record, corporate registry, binding terms, payment documentation or a directly observed public interface. Partial documentation earns partial credit. A first-party marketing statement never receives the same weight as an exact registry match.

Material contradictions, inaccessible rules and unexplained entity changes reduce the relevant criterion. We do not award points simply because a feature is common in the industry. Every positive score must be traceable to something present in the case file.

Mathematical confidence grade

Confidence uses five visible components: named-source coverage contributes 50%, documentary screenshots 20%, coverage of the six diagnostic axes 15%, a dated factual-review control 10%, and a reserved 5% for a documented funded account test. Source coverage reaches its ceiling at eight distinct, useful records and visual evidence at six useful captures. Extra links beyond the ceiling remain listed but cannot inflate confidence.

The grade thresholds are A at 94, A− at 89, B+ at 84, B at 79, B− at 74, C+ at 69, C at 64, C− at 59, D+ at 54, D at 49, D− at 44, E+ at 39, E at 34, E− at 29 and F below 29. The funded-test allocation stays at zero unless a real test occurred and can be documented.

How the five topic indices are calculated

The Withdrawal Center, KYC Center, Licence Center, Complaints Center and Authenticity Center do not reuse the overall casino score. Each publishes a documentation index derived from the relevant diagnostic axis, the number of named sources and useful documentary captures. KYC also combines withdrawal-review disclosure with a smaller identity component; authenticity combines operator continuity with domain and security evidence.

The indices measure reconstructibility only. They do not claim that a higher documentation score means faster payouts, fewer complaints, better privacy or stronger legal protection. Coverage bands are strong at 78–100, partial at 58–77, limited at 35–57 and not publicly confirmed below 35. The complete 91-row investigation index is downloadable as CSV. Thrill appears above the on-page tables as a clearly labelled featured tested alternative and is excluded from that downloadable case-file dataset.

Unknowns, conflicts and penalties

A missing public limit, undocumented manual-review ceiling or inaccessible cashier is marked unknown. Unknown is neither a positive nor proof of wrongdoing. Conflicting first-party pages are preserved as a conflict and reduce confidence until reconciled.

Market-specific information is not generalized to every account. A casino loading from one location does not prove legal eligibility. A visible licence number does not receive full credit until entity, domain, status, dates and permitted scope align.

Complaints and reputation evidence

Public reports are grouped by issue—such as delayed withdrawal, KYC after withdrawal, source-of-funds checks, bonus disputes or restricted accounts. We examine dates, supporting material, operator responses and reported resolutions.

Complaint volume alone does not establish a population-wide failure rate because the number of users and selection bias are unknown. Trustpilot star ratings are not imported into the casino score. A complaint affects the verdict only to the extent that its facts can be contextualized and corroborated.

Commercial separation and corrections

Featured alternatives may create commercial value, but they are not candidates in the investigated casino score. The verdict is completed before an alternative is selected. Sponsored outbound links are marked for search engines and do not erase documented limitations.

A change to operator, licence, eligibility, KYC policy, payment rules, withdrawal terms or scoring inputs triggers a factual review. Material changes are dated in the public correction log; cosmetic edits are not presented as research updates.

Worked calculation

Assume a case has five named sources, three useful documentary captures, all six diagnostic axes covered, a dated review and no funded test. Confidence is 5÷8×50 = 31.25; captures contribute 10; breadth 15; current review 10; funded test 0. The rounded total is 66/100, producing C.

Operating quality is separate. A case scoring 12/15 identity, 10/15 licence, 9/15 withdrawals and KYC, 10/12 payments, 6/10 complaints, 7/8 security, 6/8 games, 5/7 terms, 4/5 protection and 4/5 support totals 73/100: legitimate with limitations.